Cardiovascular Comorbidities in a United States Patient Population with Hemophilia a: A Comprehensive Chart Review

Introduction: Two previous retrospective claims database analyses (Pocoski J, et al. Haemophilia. 2014;20(4):472-478 and Humphries TJ, et al. Am J Hematol. 2016;91(5):E298-299) reported increased prevalence and earlier onset of cardiovascular (CV) comorbidities in patients with vs without hemophilia A (HEM A). Because of many known limitations of claims databases, a comprehensive chart review at a large integrated delivery network was conducted to assess differential CV comorbidities.

Aim: This study was designed to confirm the previous findings on CV risk factors and associated diseases in 2 large claims databases of male patients with HEM A in the United States.

Methods: This was a retrospective chart review study conducted at the Henry Ford Health System in patients diagnosed with HEM A (n=74) and matched Control patients (3:1) without a diagnosis of HEM A (Control, n=222). Baseline demographics, bleeding events, treatment parameters, co-existing diseases, hemophilia-associated events, and the prevalence of 12 CV risk factors and associated diseases were compared between the HEM A and Control cohorts. P values generated from a chi-square test for categorical variables and a t test for continuous variables were reported. To address the small sample size, statistical differences between the cohorts were also assessed using absolute standardized difference (SDiff), where a value ≥0.10 was considered statistically meaningful.

Results: The Control cohort was well matched to the HEM A group by age, race, healthcare payer, and study year. As expected, the prevalence of hepatitis B and C, hepatocellular carcinoma, and HIV/AIDS was much higher in the HEM A cohort. Gastrointestinal, intracranial, muscle, and joint bleeds occurred only in HEM A patients. Bleeds of various types were recorded in 35 HEM A patients vs 1 in the Control group. HEM A was severe in 52.7% of patients, moderate and mild in 10.8% each, and unknown in 25.7%. The prevalence of hypertension, diabetes, obesity, hyperlipidemia, coronary artery disease, heart failure, stroke, venous and arterial thrombosis, ventricular arrhythmias, atrial fibrillation, and chronic renal disease was numerically higher in the Control cohort, but differences were statistically significant (P≤0.05) for diabetes and hyperlipidemia only. Statistical significance using SDiff was not reached for venous and arterial thrombosis and atrial fibrillation.

Conclusions:The results of this retrospective chart review did not confirm diffuse statistically significant differences in CV comorbidities and their earlier onset in HEM A vs Controls. Reasons for the lack of confirmation are not clear but may include differences in methodology and patient populations among the studies. The Control group in this current study may have a greater medical burden than in the published studies. Our current results suggest numerically higher comorbidities in Controls for most variables. The conclusions of this study are limited by the small sample size of the hemophilia cohort and a potential selection bias associated with identification of the Control cohort.